Regulation of Iron Storage by CsrA Supports ExponentialGrowth of Escherichia coli

Christine Pourciau,a Archana Pannuri,a Anastasia Potts,a* Helen Yakhnin,b Paul Babitzke,b Tony Romeoa

aDepartment of Microbiology and Cell Science, Institute of Food and Agricultural Sciences, University of Florida, Gainesville, Florida, USAbDepartment of Biochemistry and Molecular Biology, Center for RNA Molecular Biology, The Pennsylvania State University, University Park, Pennsylvania, USA

ABSTRACT The global regulatory protein CsrA coordinates gene expression in re-sponse to physiological cues reflecting cellular stress and nutrition. CsrA binding tothe 5= segments of mRNA targets affects their translation, RNA stability, and/or tran-script elongation. Recent studies identified probable mRNA targets of CsrA that areinvolved in iron uptake and storage in Escherichia coli, suggesting an unexploredrole for CsrA in regulating iron homeostasis. Here, we assessed the impact of CsrAon iron-related gene expression, cellular iron, and growth under various iron levels.We investigated five new targets of CsrA regulation, including the genes for 4 ferri-tin or ferritin-like iron storage proteins (ISPs) and the stress-inducible Fe-S repair pro-tein, SufA. CsrA bound with high affinity and specificity to ftnB, bfr, and dps mRNAsand inhibited their translation, while it modestly activated ftnA expression. Further-more, CsrA was found to regulate cellular iron levels and support growth by repress-ing the expression of genes for ISPs, most importantly, ferritin B (FtnB) and bacterio-ferritin (Bfr). Iron starvation did not substantially affect cellular levels of CsrA or itssmall RNA (sRNA) antagonists, CsrB and CsrC. csrA disruption led to increased resis-tance to the lethal effects of H2O2 during exponential growth, consistent with a reg-ulatory role in oxidative stress resistance. We propose that during exponentialgrowth and under minimal stress, CsrA represses the deleterious expression of theISPs that function under oxidative stress and stationary-phase conditions (FtnB, Bfr,and Dps), thus ensuring that cellular iron is available to processes that are requiredfor growth.

IMPORTANCE Iron is an essential micronutrient for nearly all living organisms but istoxic in excess. Consequently, the maintenance of iron homeostasis is a critical bio-logical process, and the genes involved in this function are tightly regulated. Here,we explored a new role for the bacterial RNA binding protein CsrA in the regulationof iron homeostasis. CsrA was shown to be a key regulator of iron storage genes inEscherichia coli, with consequential effects on cellular iron levels and growth. Ourfindings establish a model in which robust CsrA activity during the exponentialphase of growth leads to repression of genes whose products sequester iron or di-vert it to unnecessary stress response processes. In so doing, CsrA supports E. coligrowth under iron-limiting laboratory conditions and may promote fitness in thecompetitive iron-limited environment of the host large intestine.

KEYWORDS CsrA, CsrB, CsrC, RNA binding proteins, bacterial growth, ferritin, ironhomeostasis, iron storage proteins, oxidative stress, posttranscriptional regulation,sRNAs, stress responses

Bacteria such as Escherichia coli have evolved complex and efficient global regula-tory systems that enable them to recognize and adapt to changing environmental

conditions, thus supporting their growth, survival, competition, and host-microbeinteractions. The carbon storage regulatory system (Csr) is one such system and plays

Citation Pourciau C, Pannuri A, Potts A,Yakhnin H, Babitzke P, Romeo T. 2019.Regulation of iron storage by CsrA supportsexponential growth of Escherichia coli. mBio10:e01034-19. https://doi.org/10.1128/mBio.01034-19.

Editor Bonnie Bassler, Princeton University

Copyright © 2019 Pourciau et al. This is anopen-access article distributed under the termsof the Creative Commons Attribution 4.0International license.

Address correspondence to Tony Romeo,tromeo@ufl.edu.

* Present address: Anastasia Potts, Takara BioUSA, Inc., Mountain View, California, USA.

Received 23 April 2019Accepted 8 July 2019Published 6 August 2019

RESEARCH ARTICLEMolecular Biology and Physiology

crossm

July/August 2019 Volume 10 Issue 4 e01034-19 ® mbio.asm.org 1

on Decem

ber 2, 2020 by guesthttp://m

bio.asm.org/

Dow

nloaded from

a critical a role in controlling numerous important cellular processes, including centralcarbon metabolism (1–3), stress response systems (4–6), biofilm formation (7), motility(8), quorum sensing (9), and virulence properties of pathogens (10). The Csr system iswidely distributed among bacteria, with homologous regulatory factors sometimesreferred to as Rsm (repressor of secondary metabolites) (11, 12).

The key component of the Csr system is CsrA, a homodimeric, sequence-specificRNA-binding protein, encoded by csrA (13–15). In general, CsrA represses the expres-sion of genes associated with stress responses and stationary-phase growth (4, 16),while activating the expression of genes that support exponential growth (16, 17). CsrAbinds to mRNA targets at sites containing a GGA motif surrounded by semiconservedsequences, with the GGA often located in the single-stranded loop of a hairpin (18, 19).The binding of CsrA to sites located in the 5= untranslated region (5= UTR) or the earlycoding region can result in altered translation (20–23), altered RNA stability (8, 24),and/or changes in RNA secondary structure that affect Rho-dependent transcriptiontermination (25).

CsrA itself is subject to multifaceted regulation (16). While csrA gene expression isregulated transcriptionally and posttranscriptionally (26), CsrA activity is controlled bythe small RNAs (sRNAs) CsrB and CsrC. These sRNAs contain multiple high-affinitybinding sites that mimic those of target mRNAs, allowing CsrB/C to bind to andsequester many CsrA dimers (27). CsrB and CsrC transcription is activated by amino acidlimitation or other stresses via ppGpp/DksA (4), metabolic end products, e.g., acetateand formate, via BarA-UvrY (12, 28), and extracytoplasmic stress via �E (6). Their RNaseE-mediated turnover is triggered by glucose, via interaction of EIIAglc with CsrD (29–31).Thus, CsrB/C sRNAs accumulate and sequester CsrA when preferred carbon resourceshave been expended, amino acids are limiting, metabolic end products have accumu-lated, and/or cells experience extracellular stress. Antagonism of CsrA by CsrB/C pro-motes the transition from glycolytic metabolism and active growth to gluconeogenesis,glycogen biosynthesis, and the formation of a stress-resistant phenotype (30).

A recent transcriptomics analysis in E. coli revealed that CsrA affects the expressionof numerous genes involved in iron uptake and metabolism and confirmed its effectson levels of four such mRNAs (19). Several of the genes were of particular interest dueto their importance in iron homeostasis and/or the large magnitude of CsrA effects (19).These effects of CsrA were suspected of occurring indirectly, as they were not con-firmed by in vivo binding of CsrA to the mRNAs using cross-linking immunoprecipita-tion sequencing (CLIP-seq). However, CLIP-seq analyses have constraints, sources ofbias (32), and are limited in sensitivity (33), potentially generating false-negative results.

Iron is an essential micronutrient for nearly all living organisms and required forfundamental processes such as respiration, central metabolism, genetic regulation, andDNA repair (34). It is the most common transition metal found in proteins, typicallywithin heme or Fe-S prosthetic groups (35, 36). Although iron is abundant in manyenvironments, ferric iron (Fe3�) predominates under aerobic conditions at neutral pH.Fe3� has poor aqueous solubility and exists as insoluble salts and oxides or bound tohost iron-binding proteins (37). Soluble ferrous iron (Fe2�) is biologically available butis rapidly oxidized to ferric iron at neutral or higher pH under aerobic conditions (35).Consequently, microbes have evolved high-affinity acquisition systems to scavengeiron (38). An added complication is that Fe2� can be cytotoxic due to its role inproducing damaging hydroxyl radicals (·OH) from H2O2 via the Fenton reaction. Thus,free intracellular iron levels are tightly controlled (34, 39). Additionally, iron availabilitycan serve as an important cue regarding the environment of the bacterium, affectingvirulence processes in many pathogens (38, 40).

Intracellular iron storage proteins (ISPs) sequester iron, providing iron reserves andprotection against toxicity (35). Three related classes of bacterial ISPs exist: archetypalferritins, heme-containing bacterioferritins, and Dps proteins (DNA protection duringstarvation). Each of these proteins is composed of identical subunits that form a roughlyspherical shell surrounding a central cavity that acts as an iron storage reservoir (34).Ferritins and bacterioferritins are composed of 24 subunits and can accommodate

Pourciau et al. ®

July/August 2019 Volume 10 Issue 4 e01034-19 mbio.asm.org 2

on Decem

ber 2, 2020 by guesthttp://m

bio.asm.org/

Dow

nloaded from

2,000 to 3,000 iron ions; Dps proteins have 12 subunits and can store �500 iron ions(34). A key feature of ISPs is the ferroxidase center, which binds two ferrous ions andoxidizes them using molecular oxygen or H2O2, forming a diferric intermediate thatmigrates to the central core for storage (35). Bacteria often possess multiple ferritin orbacterioferritin genes (34). E. coli expresses two ferritins (FtnA and FtnB), one bacte-rioferritin (Bfr), and one Dps.

Assembly of the iron-sulfur (Fe-S) clusters that serve as cofactors for enzymes is acrucial biological function, performed by Fe-S biogenesis systems (41, 42). In E. coli,most Fe-S cluster formation under nonstress conditions is by the “housekeeping” Iscpathway, encoded by the iscRSUA-hscBA-fdx (isc) operon (42, 43). Alternatively, the Sufpathway, encoded by the sufABCDSE (suf) operon, is more robust than Isc againstoxidative damage and is favored for Fe-S assembly during oxidative and nitrosativestress or when iron is limiting (44–47). The coordinated regulation of these pathwaysaccommodates the changing requirements for Fe-S cluster biogenesis, which fluctuatewith growth conditions (43, 48).

In E. coli and most bacteria, the DNA-binding ferric uptake regulator protein, Fur,controls iron metabolism (49). Fur activity depends on cellular free iron, which acts asa corepressor of transcription (40). Fur also activates gene expression via RNA poly-merase recruitment, blocking repressor access, or transcriptional repression of RyhBsRNA (36). RyhB base pairing stimulates the degradation of mRNAs encoding nones-sential iron-dependent proteins, increasing iron availability for essential processes (50).The goal of the present study was to explore the effects of CsrA on E. coli ironhomeostasis. Unlike Fur and RyhB, we show that CsrA is not responsive to ironavailability but nevertheless exerts biologically important effects on iron homeostasisand growth by regulating the expression of ISPs.

RESULTSCsrA regulates expression of iron storage genes independently of Fur. Recent

transcriptomics and other studies identified several potential mRNA targets of CsrA-mediated regulation that are involved in iron uptake and storage, suggestive of aregulatory role in iron metabolism (4, 19, 51). To further investigate this possibility,twelve of these genes were chosen for expression analysis (see Table S1 in thesupplemental material). Altogether, eleven translational =lacZ fusions and oneC-terminal 3�FLAG-tagged reporter fusion were constructed and integrated in singlecopy into the E. coli chromosome.

We assessed the expression of the following genes, which encode the indicatedproteins. We assessed four iron transport genes: (i) fhuE, an outer membrane receptorfor the xenosiderophores coprogen, ferrioxamine B, and rhodotorulic acid; (ii) fecB, theperiplasmic ferric-citrate transporter; (iii) fepA, the ferrienterobactin outer membranereceptor; and (iv) fhuA, the outer membrane transporter for ferrichrome. Four genesencoding ISPs were assessed: (i) ftnA, the primary reservoir for iron in E. coli K-12; (ii)ftnB, a ferritin-like ISP; (iii) bfr, a heme-sequestering bacterioferritin; and (iv) dps, aniron-sequestering and nonspecific DNA-binding protein that protects DNA in starvedcells. Two genes related to siderophores were investigated: (i) fes, enterobactin es-terase, which hydrolyzes the siderophore backbone to release bound iron; and (ii) entC,an isochorismate synthase involved in enterobactin biosynthesis. Lastly, the genesencoding Fur and an Fe-S cluster assembly protein, SufA, were examined.

Expression of the reporter fusions was monitored in the wild-type (WT) strainbackground MG1655 and in isogenic csrA, fur, and csrA fur mutant strains. These studiesconfirmed CsrA-dependent regulation of five iron metabolism genes (Fig. 1A to E).Translational fusions for three iron storage genes (ftnB, bfr, and dps) were stronglyrepressed by CsrA (Fig. 1A to C). The ftnB=-=lacZ fusion exhibited the most striking effect,with a 14-fold increase in the csrA mutant. Fur did not affect ftnB=-=lacZ expressionunder the conditions of this study (Fig. 1A). Expression of bfr=-=lacZ increased 7-fold inthe csrA mutant (Fig. 1B). The bfr gene has been reported to be posttranscriptionallyregulated by the Hfq-dependent sRNA RyhB, which is repressed by Fur (52). In contrast,

Regulation of Iron Storage by CsrA ®

July/August 2019 Volume 10 Issue 4 e01034-19 mbio.asm.org 3

on Decem

ber 2, 2020 by guesthttp://m

bio.asm.org/

Dow

nloaded from

we did not observe Fur-dependent regulation of the bfr=-=lacZ fusion. As the RyhBbinding site in bfr mRNA has yet to be identified, it is conceivable that our fusion didnot encompass that region. Finally, an �3-fold increase in the csrA mutant wasobserved for dps=-=lacZ, which did not respond to Fur (Fig. 1C).

Moderate CsrA-dependent effects were observed for ftnA and sufA fusions (Fig. 1Dand E). The ftnA=-=lacZ showed positive effects of CsrA on its expression, which wereretained in the Δfur mutant (Fig. 1D). Thus, CsrA has a minor role in activating ftnA geneexpression that is mediated independently of Fur. The sufA=-=lacZ fusion exhibitedmodest repression via CsrA in the fur WT and mutant backgrounds (Fig. 1E). The entC,fhuE, fepA, fes, and fhuA translational fusions were all repressed by Fur but were notregulated by CsrA (Fig. S1A to E). Expression of fur=-=lacZ was unaffected by CsrA(Fig. S1F). Altogether, these findings demonstrate that CsrA effects on iron-related geneexpression are mediated independently of Fur.

Growth phase-dependent regulation of ftnB, bfr, and dps expression by CsrA.Having identified three iron storage genes, ftnB, bfr and dps, that are repressed by CsrA,we assessed their expression during exponential phase, the transition from exponentialto stationary phase, and stationary-phase growth at 24 h. Complementation tests wereperformed using plasmid-borne csrA (pCRA16) versus the empty vector (pBR322). TheftnB, dps, and bfr fusions all exhibited growth phase-dependent regulation by CsrA,which was abolished by complementation (Fig. 2). CsrA repressed the three genesmaximally during exponential growth, with weaker effects in the transition to stationaryphase and in stationary phase. The bfr and dps fusions were not regulated by CsrAduring stationary phase (Fig. 2B and C), while ftnB repression decreased from 14-fold inthe exponential phase to �4-fold at 24 h (Fig. 2A). ftnB expression peaked during thetransition to stationary phase and remained higher in the csrA mutant in all stages ofgrowth (Fig. 2A). bfr expression levels were lowest in exponential phase, higher attransition to stationary phase, and then decreased modestly in stationary phase

FIG 1 (A to E) Effects of csrA and fur mutations on expression of translational lacZ fusions. Mean �-galactosidase activities � standard deviation weredetermined from exponential-phase cultures (OD600 of 0.5) grown in LB. Each bar shows the mean and standard deviation from four separate experiments.Statistical significance was determined using unpaired t tests and denoted as follows: ***, P � 0.001; **, P � 0.002.

Pourciau et al. ®

July/August 2019 Volume 10 Issue 4 e01034-19 mbio.asm.org 4

on Decem

ber 2, 2020 by guesthttp://m

bio.asm.org/

Dow

nloaded from

(Fig. 2B). Finally, the growth-phase expression pattern of dps reflected what was alreadyknown; Dps levels are low during exponential growth and increase upon entry tostationary phase (Fig. 2C) (53). Thus, CsrA plays a particularly important role in limitingthe expression of these three genes during exponential growth.

CsrA binds with high affinity and specificity to 5= UTRs of ftnB, dps, and bfrtranscripts. Although a CLIP-seq assay did not identify in vivo CsrA binding, previoustranscriptomics data demonstrated that CsrA affects the stability and abundance of ftnBand dps mRNA (19), suggestive of posttranscriptional regulation. To explore the pos-sible binding of CsrA to ftnB, bfr, and dps transcripts, electrophoretic gel mobility shiftassays (EMSA) were performed. CsrA exhibited high-affinity binding to all three RNAs(Fig. 3B to D). A nonlinear least-squares analysis of the data yielded apparent KD

FIG 2 Effects of csrA mutation, vector control (pBR322), and csrA complementation (pCR16) on expression of the ftnB (A), bfr (B), and dps (C) translational lacZfusions. �-Galactosidase activities � standard deviations were determined in exponential-phase (OD600 of 0.5), transition to stationary-phase (OD600 of 1.5), andovernight cultures grown in LB. Each bar shows the mean and standard deviation from four separate experiments. Statistical significance was determined usingunpaired t tests and denoted as follows: ***, P � 0.001; **, P � 0.002; *, P � 0.05.

Regulation of Iron Storage by CsrA ®

July/August 2019 Volume 10 Issue 4 e01034-19 mbio.asm.org 5

on Decem

ber 2, 2020 by guesthttp://m

bio.asm.org/

Dow

nloaded from

(equilibrium binding constant) values of 13 nM, 27 nM, and 25 nM for the ftnB, bfr, anddps RNAs, respectively, similar to those for known CsrA mRNA targets (26, 54). The bfrand dps mRNAs displayed a single shift as CsrA concentration was raised to 100 nM,while two distinct shifted species and a faint minor form were observed for ftnB at50 nM and higher, suggesting that more than one CsrA dimer may bind to this RNA.Examination of the ftnB sequence identified four potential CsrA binding sites, while thebfr and dps sequences contained three and two potential CsrA binding sites, respec-tively (Fig. 3A). Competition assays performed with specific (self) and nonspecific (phoB)unlabeled competitor transcripts confirmed that binding is specific in all cases (Fig. 3Bto D).

Taken together with reporter fusion data (Fig. 1 and 2) and transcriptomics studies(19), the high-affinity binding of CsrA to ftnB mRNA suggests that CsrA may directlyregulate expression of this gene by binding to its 5= UTR. CsrA also bound tightly to the5= UTR of dps mRNA, although the interpretation of this interaction is less clear. While

FIG 3 CsrA binding and competition reactions with ftnB (B), bfr (C), and dps (D) transcripts. (A) EMSA probe sequences with GGAsequences that are potential CsrA binding sites shown in red, translation initiation sites in green, and Shine-Dalgarno (SD) sequencesunderlined. (B to D) 5=-End-labeled transcripts (0.5 nM) were incubated with CsrA at the concentrations shown. Competition reactionswere performed in the presence of specific or nonspecific (phoB) unlabeled competitor RNAs at the concentrations shown. The positionsof free (F) and bound (B) RNA are marked with vertical bars.

Pourciau et al. ®

July/August 2019 Volume 10 Issue 4 e01034-19 mbio.asm.org 6

on Decem

ber 2, 2020 by guesthttp://m

bio.asm.org/

Dow

nloaded from

CsrA negatively affects dps RNA abundance and translation (as measured by changes inribosome occupancy), it increases dps mRNA stability (19). It is likely that CsrA bindingto the dps 5= UTR is responsible for one or more of these effects. Finally, the high-affinitybinding of CsrA to the 5= UTR of the bfr transcript, along with increased mRNAabundance in the csrA mutant (19), suggests that CsrA may directly repress bfrexpression by binding its 5= UTR.

CsrA directly represses translation of ftnB, dps, and bfr via their 5= UTRs. Toassess whether the regulatory effects of CsrA on ftnB, bfr, and dps are mediatedposttranscriptionally, without the requirement for an intermediate regulatory gene, weused the PURExpress system to measure lacZ expression from reporter fusions carriedon plasmid templates. Reporter sequences contained a T7 promoter fused to the 5= UTRof each gene and a few codons of the coding region (Fig. 3A). In all cases, with theexception of the control (pnp=-=lacZ), expression was inhibited by CsrA. Thus, CsrArepresses translation of ftnB, bfr, and dps via their 5= UTRs (Fig. 4A). The in vitrotranslational repression exhibited by the dps reporter was similar to that by the ftnBreporter. This pattern of regulation differs from what was seen in the in vivo transla-tional fusion assays, where dps demonstrated substantially less CsrA-dependent repres-sion than ftnB (Fig. 2). This difference may result from increased dps mRNA stability,which was observed previously in the csrA mutant strain (19), or it might involveindirect effects of CsrA on dps transcription in vivo.

To identify the precise CsrA interaction site(s) for the most strongly regulated gene,ftnB, RNA footprinting experiments were performed. Three regions were protectedfrom RNase T1�mediated cleavage with increasing concentrations of CsrA (Fig. 4B). Theprotected nucleotides correspond to three of the four GGA motifs, indicating that CsrAhas three authentic binding sites in the ftnB 5= UTR (Fig. 4C) (BS1, BS2, and BS3). BS3,

FIG 4 Repression of ftnB, dps, and bfr translation in vitro by CsrA and RNase T1 footprinting of ftnB RNA. (A) CsrA effects on cell-free proteinsynthesis of �-galactosidase from plasmid templates containing =lacZ fused to the 5= UTR of each mRNA target and transcribed from aT7 promoter. Relative �-galactosidase activity depicts the mean and standard deviation of activity relative to reaction mixtures lackingCsrA. (B) CsrA-ftnB RNA footprint. 5=-End-labeled ftnB RNA was treated with RNase T1 � CsrA, as shown. Partial alkaline hydrolysis (OH) andRNase T1 digestion (T1) ladders, as well as a control lane without treatment (C), are shown. Positions of the ftnB start codon (ATG) andthe Shine-Dalgarno (SD) sequence are marked. Residues protected from RNase T1 cleavage by CsrA are indicative of binding at three sites,BS1 to -3, and are shown. Numbering is with respect to the start of ftnB transcription. (C) Sequence of ftnB leader RNA. Position of thetranslation initiation codon indicated in green, GGA sequences are shown in red, and the deduced CsrA binding sites are indicated.

Regulation of Iron Storage by CsrA ®

July/August 2019 Volume 10 Issue 4 e01034-19 mbio.asm.org 7

on Decem

ber 2, 2020 by guesthttp://m

bio.asm.org/

Dow

nloaded from

which overlaps the ftnB Shine-Dalgarno (SD) sequence, exhibited the strongest protec-tion by CsrA. This finding provides a molecular mechanism for ftnB translationalrepression by CsrA, observed in vitro (Fig. 4A). The multiple protected regions areconsistent with EMSA results, which show that more than one complex is formedbetween CsrA and the ftnB 5= UTR (Fig. 3B).

Overexpression of iron storage genes in the csrA mutant increases cellular ironlevels. To determine the effect of CsrA regulation of iron storage genes on cellulariron levels, exponential-phase cultures were analyzed for iron content by induc-tively coupled plasma optical emission spectrometry (ICP-OES). Cellular iron contentwas �25% higher in the csrA mutant (5.2 � 0.15 �mol Fe/g protein) than in the WT(4.2 � 0.07 �mol Fe/g protein) and was restored to the WT level upon complemen-tation (Fig. 5A). This result is consistent with the increased expression of ftnB, bfr,and dps in the csrA mutant, which should increase its capacity for iron storage. Totest this hypothesis, deletions of the chromosomal ftnB, bfr, and dps genes wereintroduced, and cellular iron levels were measured. Iron levels in the csrA mutantapproached WT levels when FtnB and/or Bfr production was abolished (Fig. 5B),suggesting that increased iron accumulation in the csrA mutant is caused by theoverproduction of these ISPs. The dps mutation had no substantial effect on ironlevels, alone or in combination with ftnB or bfr. Notably, the increased iron in thecsrA mutant may not be biologically available. In fact, excessive iron sequestrationby FtnB and Bfr in the csrA mutant might inhibit iron-dependent cellular processes,to the detriment of cell growth.

Growth under iron-limiting conditions is compromised by csrA mutation andrestored by deletion of ftnB and bfr. To test the above hypothesis, we monitored thegrowth kinetics of WT, csrA, ftnB, bfr, and/or dps mutants under iron-replete andiron-limiting conditions in both rich and minimal media. Growth in rich medium wasassessed by tracking total cell protein of cultures grown in LB (for iron-replete condi-tions) and LB containing 400 �M 2,2=-dipyridyl (DIP) (for iron-limiting conditions).Growth in minimal medium was evaluated by measuring the optical density at 600 nm(OD600) in MOPS (morpholinepropanesulfonic acid) medium supplemented with100 �M FeSO4 (iron replete) or 0.05 �M FeSO4 (iron limiting) and 0.2% glucose (55).However, biofilm formation compromised growth measurements (A600) in MOPS me-dium. Under this condition, biofilm was apparently dependent upon the formation ofcurli fimbriae and was abolished by deletion of csgA (data not shown). Thus, all strainsexamined in MOPS medium were constructed in a csgA background.

FIG 5 Cellular iron content in E. coli strains. Cells were harvested during exponential growth (OD600 of0.5) in LB, lysed with 30% nitric acid, and iron was quantified via ICP-OES. (A) Comparison of iron levelsin WT, csrA mutant, and csrA-complemented strains. (B) Iron content of WT, csrA, and combinatorialmutants in genes for iron storage proteins. Each bar shows the mean and standard deviation from twoseparate biological experiments with paired technical replicates. Statistical significance was determinedusing unpaired t tests and is denoted as follows: **, P � 0.002; *, P � 0.05.

Pourciau et al. ®

July/August 2019 Volume 10 Issue 4 e01034-19 mbio.asm.org 8

on Decem

ber 2, 2020 by guesthttp://m

bio.asm.org/

Dow

nloaded from

Growth of the WT and csrA strains was inhibited under iron-limiting conditions inboth media, and these growth defects were particularly severe in the csrA mutant(Fig. 6; Fig. S2). To assess the contribution of each iron storage gene to the growthinhibition exhibited by the csrA mutant, deletions in the chromosomal ftnB, bfr, and dpsgenes were introduced into the WT and csrA mutant. Each strain was grown underiron-replete and iron-limiting conditions in both the rich (Fig. 6B and C; Fig. S2B and C)and minimal (Fig. 6D and E; Fig. S2D and E) media. The inactivation of ftnB, bfr, or dps

FIG 6 Growth rates (�) of exponential-phase E. coli strains in LB � 400 �M DIP and MOPS minimal medium withreplete (100 �M) or limiting (0.05 �M) FeSO4 as indicated. (A) WT, csrA mutant, and csrA-complemented strains inMOPS. (B and C) WT, csrA mutant, and ftnB/dps/bfr combinatorial mutants in LB. (D and E) WT, csrA mutant, andftnB/dps/bfr combinatorial mutants in MOPS. Each bar represents the mean and standard deviation from 4 separateexperiments. See Fig. S2 in the supplemental material for associated growth curves. Statistical significance wasdetermined using unpaired t tests. ***, P � 0.001; **, P � 0.002; *, P � 0.05.

Regulation of Iron Storage by CsrA ®

July/August 2019 Volume 10 Issue 4 e01034-19 mbio.asm.org 9

on Decem

ber 2, 2020 by guesthttp://m

bio.asm.org/

Dow

nloaded from

had no significant effect on growth of the WT strain under any condition (Fig. S2B toD). However, introduction of these knockouts into the csrA mutant led to partial orcomplete restoration of the WT growth rate in MOPS medium with limiting iron(Fig. 6E). In rich medium, the csrA mutant exhibited reduced growth only when iron waslimiting, and this growth defect was completely recovered upon inactivation of ftnBalone; dps or bfr alone did not restore growth (Fig. 6B and C). Interestingly, the csrAmutant displayed a growth defect in the minimal medium under both iron conditions,although the effects were more pronounced in the iron-limited MOPS (Fig. 6D and E).Inactivation of ftnB in the csrA mutant recovered the growth rate completely in bothiron-limited and iron-replete minimal media. The introduction of a bfr mutation to thecsrA mutant partially restored growth in iron-replete and iron-limited MOPS (Fig. 6Dand E). The dps mutation only improved growth of the csrA mutant in MOPS mediumwith limiting iron, and this effect was weak (Fig. 6D and E).

Iron availability has negligible effects on CsrA/B/C levels. A fundamental featureof Csr systems is their ability to integrate a variety of nutritional cues and signals andto interact with other regulators in diverse circuitry (10, 16, 29). Therefore, we soughtto determine whether iron availability affects levels of Csr system components. Iden-tically growing exponential-phase cultures of WT E. coli were treated with or withoutDIP (250 �M), and CsrA/B/C levels were measured. RyhB levels served as a positivecontrol for the iron starvation response (Fig. 7A) (52). CsrA was determined by Westernblotting using polyclonal anti-CsrA antibodies and normalized to RpoB (Fig. 7B). CsrAlevels were unaffected by DIP addition, indicating that CsrA expression is not regulatedin response to a change in iron availability. Northern blotting (Fig. 7C and D) showedthat CsrB levels were unaffected by iron availability, although CsrC levels droppedslightly �30 min after the addition of DIP. CsrB/C sRNAs function via CsrA, and CsrB isa more effective antagonist of CsrA than CsrC (56). Thus, the minor change in CsrClevels at 30 min might not have a significant impact on cellular physiology. In summary,it appears that iron availability has negligible effects on the key components of the Csrsystem and is not one of the cues to which the Csr system responds.

FIG 7 Effects of the addition of 250 �M 2,2=�dipyridyl (DIP) on RyhB (A), CsrA (B), CsrB (C), and CsrC (D) levels at mid-exponential-phasegrowth (OD600 of 0.5). CsrA protein was detected by Western blotting and RyhB and CsrB/C sRNAs were detected by Northern blotting.Values were normalized against RpoB or rRNA, respectively, to calculate specific levels of CsrA, CsrB, and CsrC. Fold change shows theeffects of DIP on specific CsrA, CsrB, and CsrC levels versus controls lacking DIP. Each experiment was repeated twice with essentiallyidentical results. (A) RyhB sRNA response to DIP as a positive control for iron sequestration. (B) Purified C-terminally His6-tagged CsrAprotein served as a marker for the Western blotting. (A, C, and D) RNA isolated from strains deleted for csrB and for csrC is shown in thefirst and second lanes, respectively.

Pourciau et al. ®

July/August 2019 Volume 10 Issue 4 e01034-19 mbio.asm.org 10

on Decem

ber 2, 2020 by guesthttp://m

bio.asm.org/

Dow

nloaded from

CsrA regulates resistance of E. coli against killing by H2O2. The finding that CsrArepresses sufA (Fig. 1E), which encodes a protein involved in Fe-S assembly underoxidative stress (43), suggested that CsrA may play a role in regulating resistanceagainst oxidative damage. To test this hypothesis, we exposed several isogenic strainsto H2O2 during the exponential phase of growth and monitored its effect on viability.As suspected, the csrA mutant was considerably more resistant to oxidative stress thanthe WT strain (Fig. 8). Complementation restored sensitivity of the csrA mutant to H2O2,confirming a role for CsrA in repressing oxidative stress resistance. While deleting ftnBand/or bfr did not restore H2O2 sensitivity to the csrA mutant, the loss of dps caused amarked increase in H2O2 susceptibility, resulting in greater sensitivity even than in thecsrA WT strain (Fig. 8B). Perhaps this is not surprising, given the importance of Dps inprotection against oxidative stress (57). Finally, the WT and csrA mutant strains weretested for survival against H2O2 during early stationary phase. Unexpectedly, the csrAmutant was considerably more sensitive than WT to H2O2 at this stage of growth (seeFig. S3), opposite of its response in the exponential phase (Fig. 8). These observationsare consistent with recent findings from Salmonella enterica that CsrA is a flexibleregulator whose regulon varies under different growth or physiological conditions (58).

DISCUSSION

Here, we show that CsrA regulates at least 5 genes that participate in iron storageand utilization (Fig. 1). Repression of ftnB and bfr by CsrA affects cellular iron content(Fig. 5) and is required for optimal growth under iron limitation and in minimal medium(Fig. 6). In other words, repression of ftnB and bfr by CsrA positions the Csr system toguide intracellular iron flux toward growth-promoting processes. CsrA is known tobroadly repress gene expression involved in stress responses and in the stationaryphase of growth. Furthermore, the Csr system itself does not appear to respond to ironavailability (Fig. 7). In contrast, CsrB and CsrC sRNAs are known to be expressed underconditions of stress or nutrient limitation, and by sequestering CsrA, these sRNAsdiminish CsrA-mediated regulation of gene expression when growth is no longer apriority (4, 5, 30). This regulatory arrangement no doubt allows the cellular capacity for

FIG 8 Survival of WT (MG1655) and csrA mutant strains upon exposure to H2O2. (A) WT and csrA mutant strains carryingpBR322 or plasmid pCRA16 (csrA�). (B) WT, csrA, and ftnB, dps, and bfr combinatorial mutants. Strains were grown in LBto mid-exponential phase (OD600 of 0.5) and exposed to 12.5 mM H2O2. Samples were collected 0, 15, and 30 min afteraddition of H2O2, washed, and 10-fold serially diluted. The 10�1 to 10�6 dilutions were plated and grown for 18 h on LBagar at 37°C.

Regulation of Iron Storage by CsrA ®

July/August 2019 Volume 10 Issue 4 e01034-19 mbio.asm.org 11

on Decem

ber 2, 2020 by guesthttp://m

bio.asm.org/

Dow

nloaded from

iron storage to be governed in part by nutritional and growth phase conditions thatinfluence the Csr system (29).

To our knowledge, the present studies represent the first evidence that the repres-sion of gene expression by CsrA can be necessary for robust exponential growth,although modest growth support via CsrA-mediated repression has been observed(59). While the csrA::gm allele used in this study results in a much less severe growthdefect than seen in a ΔcsrA strain, even when the csrA::gm strain was grown underlimiting iron conditions, we posit that the findings obtained with the csrA::gm mutantare nevertheless important for the following reasons. (i) Most likely, there is noenvironmental condition under which CsrA activity is absent, but there may be condi-tions under which reduced CsrA activity impacts iron requirements. (ii) In consideringthe development of CsrA inhibitors (60), complete inhibition of CsrA activity may not berequired to achieve useful therapeutic effects, because human and animal bodycompartments tend to contain little available iron (see related discussion below).

The high-affinity binding of CsrA to ftnB, dps, and bfr transcripts (Fig. 3), takentogether with in vitro translation results (Fig. 4A), indicates that CsrA represses trans-lation of ftnB, bfr, and dps mRNAs by binding to the 5= UTR of each transcript.Furthermore, RNase T1 footprinting revealed that CsrA binds to the ftnB SD sequence(Fig. 4B), which should directly impede ribosome loading. Translational repression ofhfq, glgC, cstA, and pgaA expression is similarly mediated by CsrA binding to sitesoverlapping the SD sequences (21, 54, 61). Both bfr and dps transcripts containpotential CsrA binding sites overlapping their SD sequences (Fig. 3A), which mightmediate translational repression by a similar mechanism (Fig. 4C).

The overexpression of ftnB and bfr in the csrA mutant caused excessive accumula-tion of cellular iron (Fig. 5). The csrA mutation also inhibited growth under iron-limitingconditions, apparently by altering iron bioavailability (Fig. 6). The relative magnitude ofCsrA-dependent repression of genes for ISPs (ftnB � bfr � dps) mirrored the effects ofthese genes on cellular iron content and growth (ftnB � bfr � dps) (Fig. 1A to C, 5B, and6C to E). We propose that during exponential growth, repression of the ISPs, particularlyFtnB and Bfr, prevents the sequestration of iron away from iron-dependent growthprocesses. Importantly, the regulatory effects of CsrA on these genes became weakeror were eliminated as cultures entered the stationary phase of growth (Fig. 2), duringwhich time E. coli becomes resistant to a variety of stressors (62).

The complex and dissimilar regulation of the E. coli ISPs seemingly complicates theinterpretation of their individual physiological roles (34). ftnA expression is primarilyregulated by Fur-mediated transcriptional activation, coupling FtnA levels to ironavailability (63). At high iron levels, FtnA serves as the primary reservoir in E. coli, storingup to 50% of cellular iron (64). Under low to moderate iron levels, Bfr can become theprincipal ISP (65). Notably, bfr expression is induced by stresses, including osmotic andheat stress, while ftnA expression is not (66). Likewise, ftnB transcription can be drivenby a �E-dependent promoter (67), linking FtnB production to extracytoplasmic stress, orfrom a �D-dependent promoter, as part of the Cpx regulon, connecting it to cellenvelope damage (68). dps transcription is induced under oxidative stress by thetranscriptional activator OxyR (69) and by �S during the stationary phase of growth (70).This regulatory diversity implies that individual ferritins are beneficial under differentphysiological conditions.

The strongest regulation by CsrA observed here was that of ftnB (Fig. 1A). FtnB is themajor Fe2� donor for the repair of oxidatively damaged Fe-sulfur clusters in Salmonella,although this is yet to be confirmed in E. coli (71). These findings raise the possibilitythat a primary role of CsrA in iron homeostasis is related to oxidative stress and repairresponse. CsrA also repressed bfr and dps (Fig. 1B and C), whose protein productsappear to play modest roles in iron storage in E. coli and are linked to stress responses(64). CsrA repressed expression of sufA (Fig. 1E) and negatively affected the levels ofsufABCDE operon transcripts (19), which encode proteins responsible for Fe-S clusterassembly under stress conditions. Recent research suggests that FtnB, Bfr, and Dps mayeach function to donate iron to the Suf Fe-S cluster biogenesis pathway (72). Thus, the

Pourciau et al. ®

July/August 2019 Volume 10 Issue 4 e01034-19 mbio.asm.org 12

on Decem

ber 2, 2020 by guesthttp://m

bio.asm.org/

Dow

nloaded from

present findings are consistent with a regulatory role of CsrA in the prevention andrepair of oxidative damage.

CsrA modestly activated ftnA expression, in contrast to its stronger negative effectson the genes for the other ISPs. Unlike FtnB, Bfr, and Dps, FtnA does not appear todonate iron to the Suf Fe-S cluster biogenesis pathway (72). In fact, FtnA facilitates theassembly of iron-sulfur clusters via the Isc pathway under normal physiological condi-tions (73). Although Suf Fe-S biogenesis is preferred to Isc assembly during stress, theSuf pathway fails to fully mature some Fe-S proteins and may only meet the minimalFe-S requirements for growth (43, 48, 74). As such, the activation of FtnA and repressionof FtnB, Bfr, Dps, and SufA by CsrA may assist in directing iron stores toward the Fe-Sassembly pathway that is most beneficial for exponential growth (Fig. 9). These findingssuggest an explanation for the csrA growth defect in minimal medium, even when ironis replete. The loss of CsrA activation of FtnA and repression of FtnB, Bfr, Dps, and SufAapparently result in an inefficient utilization of available iron, which is particularlydetrimental in minimal medium, where the demand for Fe-S proteins involved inbiosynthetic processes may be high (75).

Interestingly, the ferritins are not the first example of CsrA mediating opposingregulatory effects on proteins having identical or highly related activities. For example,expression of the major phosphofructokinase, PfkA, is activated by CsrA, while theminor and �s-inducible enzyme, PfkB, is repressed by CsrA (1, 19, 76). The differentallosteric responses of these enzymes to key metabolites are most likely the basis oftheir distinct regulatory patterns. To our knowledge, in such cases, CsrA consistentlyserves to repress the expression of genes associated with stationary-phase growthand/or stress responses, while tending to activate genes that are needed to supportexponential growth.

The dramatically increased resistance of the csrA mutant to H2O2 during exponentialgrowth demonstrates a regulatory role of CsrA in the resistance of E. coli to oxidativedamage (Fig. 8). This is further supported by transcriptomics studies, which foundincreased steady-state mRNA levels for genes involved in antioxidant defense in thecsrA mutant during exponential-phase growth, including the transcriptional activator ofthe superoxide response regulon, soxS, superoxide dismutases sodA and sodC, andcatalase katE (19). CsrA also copurifies with mRNAs encoding the superoxide dismutaseSodB and the catalase-peroxidase KatG (4). A recent study also provided evidence fora role for CsrA in repressing the oxidative stress response in Salmonella (58). The

FIG 9 A model for regulation of iron homeostasis by CsrA in E. coli. During exponential growth and inthe absence of stress, CsrA represses ftnB, bfr, dps, and sufA, inhibiting the delivery of iron to and theexpression of the stress-resistant Suf Fe-S cluster assembly pathway. CsrA also activates the expressionof ftnA, which donates iron to the Isc pathway. Upon stress or stationary-phase growth, CsrB/C sRNAsaccumulate and sequester CsrA, which derepresses translation of ftnB, bfr, and dps, helping to divert ironfrom the housekeeping Isc Fe-S cluster assembly pathway and toward the Suf pathway. Activation ofgene expression and iron transfer are indicated by arrows; repression is shown by a T bar.

Regulation of Iron Storage by CsrA ®

July/August 2019 Volume 10 Issue 4 e01034-19 mbio.asm.org 13

on Decem

ber 2, 2020 by guesthttp://m

bio.asm.org/

Dow

nloaded from

repression of many genes involved in oxidative stress response pathways by CsrAsuggests that ungoverned expression of these processes may be detrimental duringexponential growth and in the absence of stress, when conditions favor high CsrAactivity (30). Finally, the unexpected finding that the csrA mutant became less resistantthan the WT to H2O2 in the stationary phase of growth (see Fig. S3 in the supplementalmaterial) suggests that CsrA also performs important yet unexplained functions re-quired for oxidative stress resistance in the stationary phase.

Commensal E. coli is ubiquitous in the gastrointestinal tracts of mammals andseemingly has colonized every mammalian species on the planet (77). While muchresearch has been devoted to the role of CsrA in regulating virulence gene expressionin pathogenic Gammaproteobacteria (10), little is known about its role in commensalcolonization. E. coli inhabits the large intestine, which is primarily anoxic and nutrientlimited. In this environment, soluble free iron is a contested and often limited resource(78). In addition, reactive oxygen species are scarce in the healthy intestinal lumen,although this can change during inflammation. For commensal microbes, iron is criticalfor enzyme function and energy generation. By repressing genes with the potential tosequester iron or divert iron to unnecessary stress response processes, CsrA fosters E.coli growth under iron-limiting laboratory conditions. We propose that it may play asimilar role during colonization and survival in the large intestine.

MATERIALS AND METHODSBacterial strains and culture conditions. The bacterial strains and plasmids used in this study are

listed in Table S2 in the supplemental material. All media and medium components were prepared usingNanopure water, 18.2 M/cm (Barnstead). Bacterial strains were grown in LB broth (5 g yeast extract, 10 gBacto tryptone, and 10 g NaCl per liter double-distilled water [ddH2O]), pH 7.4, at 37°C, with shaking (250rpm), unless otherwise indicated. For growth curves under iron-limiting and iron-replete conditions,MOPS defined medium (79) with 0.2% glucose was used; FeSO4 was supplemented at the concentrationsas indicated in the figure legends. L-Amino acid stock solutions were added to MOPS medium at thefollowing final concentrations (mM): alanine (0.8), arginine (5.2), asparagine (0.4), aspartate (0.4), cysteine(0.1), glutamic acid (0.6), glutamine (0.6), glycine (0.8), histidine (0.2), isoleucine (0.4), leucine (0.8), lysine(0.4), methionine (0.2), phenylalanine (0.4), proline (0.4), serine (10.0), threonine (0.4), tryptophan (0.1),tyrosine (0.2), valine (0.6), and thiamine (0.01). When necessary, the following antibiotics were added togrowth media: ampicillin (100 �g/ml), tetracycline (15 �g/ml), gentamicin (10 �g/ml), kanamycin (50 �g/ml), and chloramphenicol (25 �g/ml). Overnight cultures were routinely used to inoculate LB broth orminimal medium unless otherwise indicated. For strains carrying a dps deletion, exponentially growingcultures were used to prepare frozen glycerol stocks and to inoculate LB broth to eliminate/minimize thetime spent under stationary-phase conditions.

Transduction with P1vir was used to introduce gene deletions and disruptions from E. coli donorstrains constructed in previous studies (31, 80) and from the Keio library (81) (Table S2). Plasmids pBR322(82) and pCRA16 (54) (csrA cloned into pBR322) were used in complementation tests. The Flp recombi-nase encoded in pCP20 (83) was used to eliminate the kanamycin resistance cassette, as required.

Considerations for strain construction. Because complete loss of CsrA activity causes severegrowth defects and genetic instability (84, 85), our experiments were performed with strains carrying amutation (csrA::gm) producing a protein that contains the first 50 of 61 amino acids of the native proteinand has residual CsrA activity (31). This csrA allele can be transduced by selection for a distally encodedgentamicin resistance (gm) marker. Strains containing fur csrA double mutations produced copiousbiofilm, which interfered with growth and reporter assays (data not shown). CsrA represses pgaABCDexpression, required for synthesis and secretion of the biofilm adhesin poly-�-1,6-N-acetyl-D-glucosamine(PGA) (54), and studies have also documented a role for Fur in biofilm formation (86, 87). Deletion ofpgaC, encoding the PGA glycosyl transferase, abolished biofilm production by the strains in LB medium.Consequently, studies in LB were performed in the pgaC mutant background.

Construction of =lacZ reporter fusions. Chromosomal translational fusions to =lacZ were con-structed using the CRIM system (88) and plasmid vector pLFT (4) and integrated at the �att site.Single-copy integrants were confirmed by PCR, as described previously (88). Constructions were per-formed as follows: �500 nucleotides (nt) of DNA upstream of and including the promoter region throughone or more codons downstream of the translational start site was amplified by PCR using the associatedprimers (see Table S3). The PCR products were gel purified, digested with PstI and BamHI, ligated intoPstI- and BamHI-digested and dephosphorylated plasmid pLFT, and electroporated into DH5� �pir cells.The fusion sequences were verified, and plasmids were isolated and integrated into the �att site of strainMG1655 ΔlacZ using the helper plasmid pFINT (4).

�-Galactosidase assay. Strains containing =lacZ fusions were grown at 37°C in Luria-Bertani (LB)broth. Tetracycline (15 �g/ml) was used to ensure plasmid maintenance in strains bearing pBR322 orpCRA16. Cells were harvested at various times and �-galactosidase activity was measured as describedpreviously (4). Total cell protein was measured following precipitation with 10% trichloroacetic acid,

Pourciau et al. ®

July/August 2019 Volume 10 Issue 4 e01034-19 mbio.asm.org 14

on Decem

ber 2, 2020 by guesthttp://m

bio.asm.org/

Dow

nloaded from

using the bicinchoninic acid (BCA) assay (Pierce Biotechnology) with bovine serum albumin as theprotein standard.

Electrophoretic gel mobility shift assays for RNA binding. The binding of CsrA to ftnB, dps, and bfrtranscripts was determined by EMSA using in vitro-synthesized ftnB, dps, and bfr transcripts (MAXIscriptSP6/MEGAshortscript T7 kits; Ambion) and recombinant CsrA-His6 (14). The template DNAs for in vitrotranscription of ftnB and bfr were generated by PCR from MG1655 genomic DNA, using oligonucleotidepairs ftnB SP6 fwd EMSA/ftnB SP6 rev EMSA and bfr T7 fwd EMSA/bfr T7 rev EMSA. The template DNA forin vitro transcription of dps was generated by annealing the oligonucleotide pair dps T7 EMSA/dps T7EMSA comp. RNA was synthesized from the ftnB (127 nt, consisting of 127 nt of the 5= UTR), dps (50 nt,consisting of 39 nt of the 5= UTR and 11 nt of the coding region), and bfr (81 nt, consisting of the 23 nt5= UTR and 58 nt of the coding region) templates in vitro using the MEGAshortscript kit (Ambion). Theresulting transcripts were purified via denaturing polyacrylamide gel electrophoresis (PAGE) followed byphenol-chloroform extraction and ethanol precipitation. Transcripts were treated with Antarctic phos-phatase (NEB) and radiolabeled at the 5= end using [�-32P]ATP and T4 polynucleotide kinase. Bindingreaction mixtures contained 0.5 nM RNA, 10 mM MgCl2, 100 mM KCl, 32.5 ng total yeast RNA, 20 mMdithiothreitol (DTT), 7.5% glycerol, 4 U SUPERasin (Ambion), and various concentrations of recombinantCsrA and were incubated at 37°C for 30 min. Reaction mixtures were separated on 9% native polyacryl-amide gels (for ftnB mRNA) and 12% native polyacrylamide gels (for dps and bfr mRNAs) with 1�Tris-borate-EDTA (TBE) buffer. Competition assays were performed in the presence or absence ofunlabeled specific (self) and nonspecific (phoB) RNA competitors using the minimum CsrA concentrationrequired for a full shift. Labeled RNA was analyzed using a phosphorimager equipped with Quantity Onesoftware (Bio-Rad), as previously described (4). The apparent equilibrium binding constant (KD) forCsrA-RNA complex formation was calculated according to a previously described cooperative bindingequation (14).

Footprint assay. CsrA-ftnB RNA footprint assays were performed according to a published procedure(24). ftnB RNA (nt �1 to �152) was synthesized with the RNAMaxx kit (Agilent Technologies) usingPCR-generated DNA templates. Gel-purified RNA was dephosphorylated and then 5= end labeled usingT4 polynucleotide kinase (New England BioLabs) and [�-32P]ATP (7,000 Ci/mmol). Labeled RNAs wererenatured by heating for 1 min at 90°C followed by slow cooling to room temperature. Binding reactionmixtures (10 �l) contained 2 nM labeled RNA, 10 mM Tris-HCl (pH 7.5), 10 mM MgCl2, 100 mM KCl, 40 ngof yeast RNA, 7.5% glycerol, 0.1 mg/ml xylene cyanol, and various concentrations of purified CsrA-His6.After a 30-min incubation at 37°C to allow for CsrA-RNA complex formation, RNase T1 (0.016 U) wasadded and the incubation was continued for 15 min at 37°C. The reactions were stopped by adding 10 �lof stop solution (95% formamide, 0.025% SDS, 20 mM EDTA, 0.025% bromophenol blue, 0.025% xylenecyanol). Samples were heated for 5 min at 90°C and fractionated through standard 6% (vol/vol)polyacrylamide-8 M urea sequencing gels. Cleaved patterns were examined using a Typhoon 8600variable mode imager.

Coupled transcription-translation assay. In vitro coupled transcription-translation assays usingPURExpress (New England BioLabs) were performed according to a published procedure (89). PlasmidpYH333 contains a T7 promoter driving transcription of the bfr translational fusion (nt �1 to �50 relativeto the bfr transcriptional start site). Plasmid pYH334 contains a T7 promoter driving transcription of thedps translational fusion (nt �1 to �81 relative to the dps transcriptional start site). Plasmid pYH336contains a T7 promoter driving transcription of the ftnB translational fusion (nt �1 to �152 relative tothe ftnB transcriptional start site). These plasmids were used as the templates for coupled transcription-translation reactions using the PURExpress in vitro protein synthesis kit, according to the manufacturer’sinstructions. Each 6.7-�l reaction mixture contained 250 ng of plasmid DNA template, various concen-trations of purified CsrA-H6, 1 U of RNasin (Promega), 2.5 mM DTT, 2.7 �l of solution A, and 2 �l ofsolution B. The mixtures were incubated for 2.5 h at 37°C, and �-galactosidase activity was determinedaccording to the manufacturer’s instructions.

Cellular iron measurement. Total cellular Fe concentrations were measured by inductively coupledplasma optical emission spectrometry (ICP-OES) at the University of Florida Institute of Food andAgricultural Sciences Analytical Services Laboratories. LB medium (500 ml) was inoculated with overnightcultures and incubated at 37°C with shaking. Two hundred fifty milliliters of exponential-phase cultures(OD600 of 0.5) were harvested by centrifugation and washed three times in cold phosphate-bufferedsaline (PBS). The resulting bacterial pellets were resuspended in 1 ml 35% HNO3 (trace-metal grade)treated at 95°C and diluted 1:10 with Invitrogen UltraPure distilled water. Fe concentrations werenormalized to total protein. For quantification of iron in the medium, 2 ml trace-metal-grade 35% HNO3

was added to 18 ml medium for ICP-OES analysis. Glassware for these experiments was pretreatedovernight to remove trace metal contamination with 20% trace-metal-grade HNO3 and multiple rinseswith Nanopure water.

Growth kinetics assay. Growth under Fe-limiting and -replete conditions was monitored by OD600

measurements in MOPS minimal medium (see “Bacterial strains and culture conditions”) or by BCAprotein assay in Luria-Bertani (LB) medium. All medium components were filter sterilized. Prior to use,labware was treated overnight in a 20% trace-metal-grade HNO3 and rinsed multiple times withNanopure water. ICP-OES analysis indicated that final Fe concentration of the MOPS medium before Fesupplementation was below 0.01 �M. For growth experiments performed in MOPS, overnight cultureswere grown in MOPS with 0.05 �M FeSO4 and were diluted 1:100 (except for cultures carrying the dpsmutation, which were inoculated directly from �80°C glycerol stocks) and grown to exponential phasein MOPS with 0.05 or 100 �M FeSO4. Growth experiments in LB were started from overnight cultures inLB, which were diluted 1:100 and grown to exponential phase in LB. The growth curves in all media were

Regulation of Iron Storage by CsrA ®

July/August 2019 Volume 10 Issue 4 e01034-19 mbio.asm.org 15

on Decem

ber 2, 2020 by guesthttp://m

bio.asm.org/

Dow

nloaded from

started at an OD600 of 0.01 from the exponentially growing cultures. Total cell protein was measuredfollowing precipitation with 10% trichloroacetic acid, using the BCA assay (Pierce Biotechnology) withbovine serum albumin as the protein standard. The growth rate constant (�) was calculated from theexponential phase of growth: � 2.303(logOD2 � logOD1)/(t2 – t1).

Western blotting of CsrA and FecB-3�FLAG protein. Samples for CsrA immunoblotting wereharvested at intervals after the addition of 250 �M iron chelator 2,2=�dipyridyl. Samples for FecB-3�FLAGblots were harvested at an OD600 of 0.05 from LB supplemented with 1 mM sodium citrate to induceexpression. Cells were centrifuged and resuspended in 2� sample buffer (4% [wt/vol] SDS, 0.16 M Tris,1.5% [vol/vol] �-mercaptoethanol, 20% [vol/vol] glycerol, 0.02% [wt/vol] bromophenol blue, pH 6.0),normalized by the OD, and lysed by sonication and boiling. Samples were separated by SDS-PAGE,transferred to 0.2 �M polyvinylidene difluoride (PVDF) membranes, and detected using polyclonalanti-CsrA or monoclonal anti-FLAG for FecB-3�FLAG, or anti-RpoB antibodies, as described previously(90).

Northern blotting. Samples were harvested and total cellular RNA was isolated using the RNeasyminikit (Qiagen). Total RNA was mixed with 2 volumes of loading buffer (50% [vol/vol] deionizedformamide, 6% [vol/vol] formaldehyde, 1� MOPS [20 mM], 5 mM sodium acetate [NaOAc], 2 mM EDTA[pH 7.0], 10% [vol/vol] glycerol, 0.05% [wt/vol] bromophenol blue, 0.01% [wt/vol] ethidium bromide),denatured by heating at 95°C for 5 min, placed on ice, and separated by electrophoresis on a 7 Murea-5% polyacrylamide gel. RNA was transferred and fixed to a positively charged nylon membrane(Roche) and hybridized with digoxigenin (DIG)-labeled antisense CsrB, CsrC, or RyhB RNA probes aspreviously described (5). Transferred rRNA served as a loading control and was stained with methyleneblue, imaged using a Gel-Doc, and signal intensity was quantified using Quantity One software. CsrB orCsrC RNAs signals were captured with a ChemiDoc XRS� system (Bio-Rad, Hercules, CA).

Oxidative stress assay. WT and mutant strains were grown in LB to mid-exponential (OD600 of 0.5)or early stationary (OD600 of 2.5) phase and exposed to H2O2 (12.5 mM and 50.0 mM, respectively). After0, 15, and 30 min, samples were collected, washed 3 times in pH 7.4 PBS, and 10-fold serially diluted. Cellswere plated on LB and grown for 18 h at 37°C before imaging the resulting colonies.

SUPPLEMENTAL MATERIALSupplemental material for this article may be found at https://doi.org/10.1128/mBio

.01034-19.FIG S1, TIF file, 2.9 MB.FIG S2, TIF file, 2.9 MB.FIG S3, TIF file, 2.2 MB.TABLE S1, DOCX file, 0.1 MB.TABLE S2, DOCX file, 0.1 MB.TABLE S3, DOCX file, 0.1 MB.

ACKNOWLEDGMENTSThis project was funded by NIH R01 grant GM059969 to Tony Romeo and Paul

Babitzke and by the National Science Foundation Graduate Research Fellowship Pro-gram (DGE-1315138 to Anastasia Potts).

The funders had no role in study design, data collection and analysis, decision topublish, or preparation of the manuscript.

REFERENCES1. Sabnis NA, Yang H, Romeo T. 1995. Pleiotropic regulation of central

carbohydrate metabolism in Escherichia coli via the gene csrA. J BiolChem 270:29096 –29104. https://doi.org/10.1074/jbc.270.49.29096.

2. Morin M, Ropers D, Letisse F, Laguerre S, Portais JC, Cocaign-BousquetM, Enjalbert B. 2016. The post-transcriptional regulatory system CSRcontrols the balance of metabolic pools in upper glycolysis of Escherichiacoli. Mol Microbiol 100:686 –700. https://doi.org/10.1111/mmi.13343.

3. Revelles O, Millard P, Nougayrède JP, Dobrindt U, Oswald E, Létisse F,Portais JC. 2013. The carbon storage regulator (Csr) system exerts anutrient-specific control over central metabolism in Escherichia coli strainNissle 1917. PLoS One 8:e66386. https://doi.org/10.1371/journal.pone.0066386.

4. Edwards AN, Patterson-Fortin LM, Vakulskas CA, Mercante JW, PotrykusK, Vinella D, Camacho MI, Fields JA, Thompson SA, Georgellis D, CashelM, Babitzke P, Romeo T. 2011. Circuitry linking the Csr and stringentresponse global regulatory systems. Mol Microbiol 80:1561–1580.https://doi.org/10.1111/j.1365-2958.2011.07663.x.

5. Pannuri A, Vakulskas CA, Zere T, McGibbon LC, Edwards AN, Georgellis D,Babitzke P, Romeo T. 2016. Circuitry linking the catabolite repression and

Csr global regulatory systems of Escherichia coli. J Bacteriol 198:3000 –3015. https://doi.org/10.1128/JB.00454-16.

6. Yakhnin H, Aichele R, Ades SE, Romeo T, Babitzke P. 2017. Circuitrylinking the global Csr and �E-dependent cell envelope stress responsesystems. J Bacteriol 199:e00484-17. https://doi.org/10.1128/JB.00484-17.

7. Jackson DW, Suzuki K, Oakford L, Simecka JW, Hart ME, Romeo T. 2002.Biofilm formation and dispersal under the influence of the global reg-ulator CsrA of Escherichia coli. J Bacteriol 184:290 –301. https://doi.org/10.1128/jb.184.1.290-301.2002.

8. Wei BL, Brun-Zinkernagel AM, Simecka JW, Prüss BM, Babitzke P, RomeoT. 2001. Positive regulation of motility and flhDC expression by theRNA-binding protein CsrA of Escherichia coli. Mol Microbiol 40:245–256.https://doi.org/10.1046/j.1365-2958.2001.02380.x.

9. Yakhnin H, Baker CS, Berezin I, Evangelista MA, Rassin A, Romeo T,Babitzke P. 2011. CsrA represses translation of sdiA, which encodes theN-acylhomoserine-L-lactone receptor of Escherichia coli, by binding ex-clusively within the coding region of sdiA mRNA. J Bacteriol 193:6162– 6170. https://doi.org/10.1128/JB.05975-11.

10. Vakulskas CA, Potts AH, Babitzke P, Ahmer BM, Romeo T. 2015. Regula-

Pourciau et al. ®

July/August 2019 Volume 10 Issue 4 e01034-19 mbio.asm.org 16

on Decem

ber 2, 2020 by guesthttp://m

bio.asm.org/

Dow

nloaded from

tion of bacterial virulence by Csr (Rsm) systems. Microbiol Mol Biol Rev79:193–224. https://doi.org/10.1128/MMBR.00052-14.

11. White D, Hart ME, Romeo T. 1996. Phylogenetic distribution of the globalregulatory gene csrA among eubacteria. Gene 182:221–223. https://doi.org/10.1016/s0378-1119(96)00547-1.

12. Zere TR, Vakulskas CA, Leng Y, Pannuri A, Potts AH, Dias R, Tang D,Kolaczkowski B, Georgellis D, Ahmer BM, Romeo T. 2015. Genomictargets and features of BarA-UvrY (-SirA) signal transduction systems.PLoS One 10:e0145035. https://doi.org/10.1371/journal.pone.0145035.

13. Romeo T, Gong M, Liu MY, Brun-Zinkernagel AM. 1993. Identificationand molecular characterization of csrA, a pleiotropic gene from Esche-richia coli that affects glycogen biosynthesis, gluconeogenesis, cell size,and surface properties. J Bacteriol 175:4744 – 4755. https://doi.org/10.1128/jb.175.15.4744-4755.1993.

14. Mercante J, Suzuki K, Cheng X, Babitzke P, Romeo T. 2006. Comprehen-sive alanine-scanning mutagenesis of Escherichia coli CsrA defines twosubdomains of critical functional importance. J Biol Chem 281:31832–31842. https://doi.org/10.1074/jbc.M606057200.

15. Schubert M, Lapouge K, Duss O, Oberstrass FC, Jelesarov I, Haas D, AllainFH. 2007. Molecular basis of messenger RNA recognition by the specificbacterial repressing clamp RsmA/CsrA. Nat Struct Mol Biol 14:807– 813.https://doi.org/10.1038/nsmb1285.

16. Romeo T, Vakulskas CA, Babitzke P. 2013. Post-transcriptional regulationon a global scale: form and function of Csr/Rsm systems. Environ Micro-biol 15:313–324. https://doi.org/10.1111/j.1462-2920.2012.02794.x.

17. Romeo T. 1998. Global regulation by the small RNA-binding protein CsrAand the non-coding RNA molecule CsrB. Mol Microbiol 29:1321–1330.https://doi.org/10.1046/j.1365-2958.1998.01021.x.

18. Dubey AK, Baker CS, Romeo T, Babitzke P. 2005. RNA sequence andsecondary structure participate in high-affinity CsrA-RNA interaction.RNA 11:1579 –1587. https://doi.org/10.1261/rna.2990205.

19. Potts AH, Vakulskas CA, Pannuri A, Yakhnin H, Babitzke P, Romeo T. 2017.Global role of the bacterial post-transcriptional regulator CsrA revealedby integrated transcriptomics. Nat Commun 8:1596. https://doi.org/10.1038/s41467-017-01613-1.

20. Yakhnin H, Pandit P, Petty TJ, Baker CS, Romeo T, Babitzke P. 2007. CsrAof Bacillus subtilis regulates translation initiation of the gene encodingthe flagellin protein (hag) by blocking ribosome binding. Mol Microbiol64:1605–1620. https://doi.org/10.1111/j.1365-2958.2007.05765.x.

21. Baker CS, Eory LA, Yakhnin H, Mercante J, Romeo T, Babitzke P. 2007.CsrA inhibits translation initiation of Escherichia coli hfq by binding to asingle site overlapping the Shine-Dalgarno sequence. J Bacteriol 189:5472–5481. https://doi.org/10.1128/JB.00529-07.

22. Patterson-Fortin LM, Vakulskas CA, Yakhnin H, Babitzke P, Romeo T. 2013.Dual posttranscriptional regulation via a cofactor-responsive mRNA leader.J Mol Biol 425:3662–3677. https://doi.org/10.1016/j.jmb.2012.12.010.

23. Irie Y, Starkey M, Edwards AN, Wozniak DJ, Romeo T, Parsek MR. 2010.Pseudomonas aeruginosa biofilm matrix polysaccharide Psl is regulatedtranscriptionally by RpoS and post-transcriptionally by RsmA. Mol Mi-crobiol 78:158 –172. https://doi.org/10.1111/j.1365-2958.2010.07320.x.

24. Yakhnin AV, Baker CS, Vakulskas CA, Yakhnin H, Berezin I, Romeo T,Babitzke P. 2013. CsrA activates flhDC expression by protecting flhDCmRNA from RNase E-mediated cleavage. Mol Microbiol 87:851– 866.https://doi.org/10.1111/mmi.12136.

25. Figueroa-Bossi N, Schwartz A, Guillemardet B, D’Heygère F, Bossi L,Boudvillain M. 2014. RNA remodeling by bacterial global regulator CsrApromotes Rho-dependent transcription termination. Genes Dev 28:1239 –1251. https://doi.org/10.1101/gad.240192.114.

26. Yakhnin H, Yakhnin AV, Baker CS, Sineva E, Berezin I, Romeo T, BabitzkeP. 2011. Complex regulation of the global regulatory gene csrA: CsrA-mediated translational repression, transcription from five promoters byE�70 and E�S, and indirect transcriptional activation by CsrA. Mol Micro-biol 81:689 –704. https://doi.org/10.1111/j.1365-2958.2011.07723.x.

27. Babitzke P, Romeo T. 2007. CsrB sRNA family: sequestration of RNA-binding regulatory proteins. Curr Opin Microbiol 10:156 –163. https://doi.org/10.1016/j.mib.2007.03.007.

28. Chavez RG, Alvarez AF, Romeo T, Georgellis D. 2010. The physiologicalstimulus for the BarA sensor kinase. J Bacteriol 192:2009 –2012. https://doi.org/10.1128/JB.01685-09.

29. Romeo T, Babitzke P. 2018. Global regulation by CsrA and its RNAantagonists. Microbiol Spectr 6:RWR-0009-2017. https://doi.org/10.1128/microbiolspec.RWR-0009-2017.

30. Leng Y, Vakulskas CA, Zere TR, Pickering BS, Watnick PI, Babitzke P,Romeo T. 2016. Regulation of CsrB/C sRNA decay by EIIAGlc of the

phosphoenolpyruvate: carbohydrate phosphotransferase system. MolMicrobiol 99:627– 639. https://doi.org/10.1111/mmi.13259.

31. Vakulskas CA, Leng Y, Abe H, Amaki T, Okayama A, Babitzke P, Suzuki K,Romeo T. 2016. Antagonistic control of the turnover pathway for theglobal regulatory sRNA CsrB by the CsrA and CsrD proteins. NucleicAcids Res 44:7896 –7910. https://doi.org/10.1093/nar/gkw484.

32. Darnell RB. 2010. HITS-CLIP: panoramic views of protein-RNA regulationin living cells. Wiley Interdiscip Rev RNA 1:266 –286. https://doi.org/10.1002/wrna.31.

33. Wheeler EC, Van Nostrand EL, Yeo GW. 2018. Advances and challengesin the detection of transcriptome-wide protein-RNA interactions. WileyInterdiscip Rev RNA 9:e1436. https://doi.org/10.1002/wrna.1436.

34. Andrews SC, Robinson AK, Rodríguez-Quiñones F. 2003. Bacterial ironhomeostasis. FEMS Microbiol Rev 27:215–237. https://doi.org/10.1016/S0168-6445(03)00055-X.

35. Andrews S, Norton I, Salunkhe AS, Goodluck H, Aly WS, Mourad-Agha H,Cornelis P. 2013. Control of iron metabolism in bacteria. Met Ions Life Sci12:203–239. https://doi.org/10.1007/978-94-007-5561-1_7.

36. Salvail H, Massé E. 2012. Regulating iron storage and metabolism withRNA: an overview of posttranscriptional controls of intracellular ironhomeostasis. Wiley Interdiscip Rev RNA 3:26 –36. https://doi.org/10.1002/wrna.102.

37. Braun V. 2001. Iron uptake mechanisms and their regulation in patho-genic bacteria. Int J Med Microbiol 291:67–79. https://doi.org/10.1078/1438-4221-00103.

38. Skaar EP. 2010. The battle for iron between bacterial pathogens andtheir vertebrate hosts. PLoS Pathog 6:e1000949. https://doi.org/10.1371/journal.ppat.1000949.

39. Frawley ER, Fang FC. 2014. The ins and outs of bacterial iron metabolism.Mol Microbiol 93:609 – 616. https://doi.org/10.1111/mmi.12709.

40. Porcheron G, Dozois CM. 2015. Interplay between iron homeostasis andvirulence: Fur and RyhB as major regulators of bacterial pathogenicity.Vet Microbiol 179:2–14. https://doi.org/10.1016/j.vetmic.2015.03.024.

41. Lill R. 2009. Function and biogenesis of iron-sulphur proteins. Nature460:831– 838. https://doi.org/10.1038/nature08301.

42. Roche B, Aussel L, Ezraty B, Mandin P, Py B, Barras F. 2013. Iron/sulfurproteins biogenesis in prokaryotes: formation, regulation and diversity.Biochim Biophys Acta 1827:455– 469. https://doi.org/10.1016/j.bbabio.2012.12.010.

43. Mettert EL, Kiley PJ. 2015. How is Fe-S cluster formation regulated? AnnuRev Microbiol 69:505–526. https://doi.org/10.1146/annurev-micro-091014-104457.

44. Jang S, Imlay JA. 2010. Hydrogen peroxide inactivates the Escherichia coliIsc iron-sulphur assembly system, and OxyR induces the Suf system tocompensate. Mol Microbiol 78:1448 –1467. https://doi.org/10.1111/j.1365-2958.2010.07418.x.

45. Outten FW, Djaman O, Storz G. 2004. A suf operon requirement for Fe-Scluster assembly during iron starvation in Escherichia coli. Mol Microbiol52:861– 872. https://doi.org/10.1111/j.1365-2958.2004.04025.x.

46. Justino MC, Vicente JB, Teixeira M, Saraiva LM. 2005. New genes impli-cated in the protection of anaerobically grown Escherichia coli againstnitric oxide. J Biol Chem 280:2636 –2643. https://doi.org/10.1074/jbc.M411070200.

47. Dai Y, Outten FW. 2012. The E. coli SufS-SufE sulfur transfer system ismore resistant to oxidative stress than IscS-IscU. FEBS Lett 586:4016 – 4022. https://doi.org/10.1016/j.febslet.2012.10.001.

48. Mettert EL, Kiley PJ. 2014. Coordinate regulation of the Suf and Isc Fe-Scluster biogenesis pathways by IscR is essential for viability of Escherichiacoli. J Bacteriol 196:4315– 4323. https://doi.org/10.1128/JB.01975-14.

49. Fillat MF. 2014. The FUR (ferric uptake regulator) superfamily: diversityand versatility of key transcriptional regulators. Arch Biochem Biophys546:41–52. https://doi.org/10.1016/j.abb.2014.01.029.

50. Jacques JF, Jang S, Prévost K, Desnoyers G, Desmarais M, Imlay J, MasséE. 2006. RyhB small RNA modulates the free intracellular iron pool andis essential for normal growth during iron limitation in Escherichia coli.Mol Microbiol 62:1181–1190. https://doi.org/10.1111/j.1365-2958.2006.05439.x.

51. Kulkarni PR, Jia T, Kuehne SA, Kerkering TM, Morris ER, Searle MS, HeebS, Rao J, Kulkarni RV. 2014. A sequence-based approach for prediction ofCsrA/RsmA targets in bacteria with experimental validation in Pseu-domonas aeruginosa. Nucleic Acids Res 42:6811– 6825. https://doi.org/10.1093/nar/gku309.

52. Massé E, Gottesman S. 2002. A small RNA regulates the expression of

Regulation of Iron Storage by CsrA ®

July/August 2019 Volume 10 Issue 4 e01034-19 mbio.asm.org 17

on Decem

ber 2, 2020 by guesthttp://m

bio.asm.org/

Dow

nloaded from

genes involved in iron metabolism in Escherichia coli. Proc Natl Acad SciU S A 99:4620 – 4625. https://doi.org/10.1073/pnas.032066599.

53. Grainger DC, Goldberg MD, Lee DJ, Busby SJ. 2008. Selective repressionby Fis and H-NS at the Escherichia coli dps promoter. Mol Microbiol68:1366 –1377. https://doi.org/10.1111/j.1365-2958.2008.06253.x.

54. Wang X, Dubey AK, Suzuki K, Baker CS, Babitzke P, Romeo T. 2005. CsrApost-transcriptionally represses pgaABCD, responsible for synthesis of abiofilm polysaccharide adhesin of Escherichia coli. Mol Microbiol 56:1648 –1663. https://doi.org/10.1111/j.1365-2958.2005.04648.x.

55. Hartmann A, Braun V. 1981. Iron uptake and iron limited growth ofEscherichia coli K-12. Arch Microbiol 130:353–356. https://doi.org/10.1007/BF00414599.

56. Weilbacher T, Suzuki K, Dubey AK, Wang X, Gudapaty S, Morozov I, BakerCS, Georgellis D, Babitzke P, Romeo T. 2003. A novel sRNA component ofthe carbon storage regulatory system of Escherichia coli. Mol Microbiol48:657– 670. https://doi.org/10.1046/j.1365-2958.2003.03459.x.

57. Calhoun LN, Kwon YM. 2011. Structure, function and regulation of theDNA-binding protein Dps and its role in acid and oxidative stressresistance in Escherichia coli: a review. J Appl Microbiol 110:375–386.https://doi.org/10.1111/j.1365-2672.2010.04890.x.

58. Potts AH, Guo Y, Ahmer BMM, Romeo T. 2019. Role of CsrA in stressresponses and metabolism important for Salmonella virulence revealedby integrated transcriptomics. PLoS One 14:e0211430. https://doi.org/10.1371/journal.pone.0211430.

59. Morin M, Ropers D, Cinquemani E, Portais JC, Enjalbert B, Cocaign-Bousquet M. 2017. The Csr system regulates Escherichia coli fitness bycontrolling glycogen accumulation and energy levels. mBio 8:e01628-17.https://doi.org/10.1128/mBio.01628-17.

60. Maurer CK, Fruth M, Empting M, Avrutina O, Hoßmann J, Nadmid S,Gorges J, Herrmann J, Kazmaier U, Dersch P, Müller R, Hartmann RW.2016. Discovery of the first small-molecule CsrA-RNA interaction inhib-itors using biophysical screening technologies. Future Med Chem8:931–947. https://doi.org/10.4155/fmc-2016-0033.

61. Dubey AK, Baker CS, Suzuki K, Jones AD, Pandit P, Romeo T, Babitzke P.2003. CsrA regulates translation of the Escherichia coli carbon starvationgene, cstA, by blocking ribosome access to the cstA transcript. J Bacteriol185:4450 – 4460. https://doi.org/10.1128/jb.185.15.4450-4460.2003.

62. Battesti A, Majdalani N, Gottesman S. 2011. The RpoS-mediated generalstress response in Escherichia coli. Annu Rev Microbiol 65:189 –213.https://doi.org/10.1146/annurev-micro-090110-102946.

63. Nandal A, Huggins CC, Woodhall MR, McHugh J, Rodríguez-Quiñones F,Quail MA, Guest JR, Andrews SC. 2010. Induction of the ferritin gene(ftnA) of Escherichia coli by Fe2�-Fur is mediated by reversal of H-NSsilencing and is RyhB independent. Mol Microbiol 75:637– 657. https://doi.org/10.1111/j.1365-2958.2009.06977.x.

64. Abdul-Tehrani H, Hudson AJ, Chang YS, Timms AR, Hawkins C, WilliamsJM, Harrison PM, Guest JR, Andrews SC. 1999. Ferritin mutants of Esch-erichia coli are iron deficient and growth impaired, and fur mutants areiron deficient. J Bacteriol 181:1415–1428.

65. Outten W, Rouault T (ed). 2014. Iron-sulfur clusters in chemistry andbiology. Walter de Gruyter GmbH, Berlin, Germany.

66. Gunasekera TS, Csonka LN, Paliy O. 2008. Genome-wide transcriptionalresponses of Escherichia coli K-12 to continuous osmotic and heatstresses. J Bacteriol 190:3712–3720. https://doi.org/10.1128/JB.01990-07.

67. Rhodius VA, Suh WC, Nonaka G, West J, Gross CA. 2006. Conserved andvariable functions of the �E stress response in related genomes. PLoSBiol 4:e2. https://doi.org/10.1371/journal.pbio.0040002.

68. Yamamoto K, Ishihama A. 2006. Characterization of copper-induciblepromoters regulated by CpxA/CpxR in Escherichia coli. Biosci BiotechnolBiochem 70:1688 –1695. https://doi.org/10.1271/bbb.60024.

69. Altuvia S, Almirón M, Huisman G, Kolter R, Storz G. 1994. The dpspromoter is activated by OxyR during growth and by IHF and �S instationary phase. Mol Microbiol 13:265–272. https://doi.org/10.1111/j.1365-2958.1994.tb00421.x.

70. Singh SS, Typas A, Hengge R, Grainger DC. 2011. Escherichia coli �70

senses sequence and conformation of the promoter spacer region.Nucleic Acids Res 39:5109 –5118. https://doi.org/10.1093/nar/gkr080.

71. Velayudhan J, Castor M, Richardson A, Main-Hester KL, Fang FC. 2007.The role of ferritins in the physiology of Salmonella enterica sv.Typhimurium: a unique role for ferritin B in iron-sulphur cluster repair

and virulence. Mol Microbiol 63:1495–1507. https://doi.org/10.1111/j.1365-2958.2007.05600.x.

72. Bolaji N. 2017. Characterization of the SUF FE-S pathway in Escherichiacoli. University of South Carolina, Columbia, SC.

73. Bitoun JP, Wu G, Ding H. 2008. Escherichia coli FtnA acts as an iron buffer forre-assembly of iron-sulfur clusters in response to hydrogen peroxide stress.Biometals 21:693–703. https://doi.org/10.1007/s10534-008-9154-7.

74. Ezraty B, Vergnes A, Banzhaf M, Duverger Y, Huguenot A, Brochado AR, SuSY, Espinosa L, Loiseau L, Py B, Typas A, Barras F. 2013. Fe-S clusterbiosynthesis controls uptake of aminoglycosides in a ROS-less death path-way. Science 340:1583–1587. https://doi.org/10.1126/science.1238328.

75. Tao H, Bausch C, Richmond C, Blattner FR, Conway T. 1999. Functionalgenomics: expression analysis of Escherichia coli growing on minimaland rich media. J Bacteriol 181:6425– 6440.

76. Lacour S, Landini P. 2004. �S-dependent gene expression at the onset ofstationary phase in Escherichia coli: function of �S-dependent genes andidentification of their promoter sequences. J Bacteriol 186:7186 –7195.https://doi.org/10.1128/JB.186.21.7186-7195.2004.

77. Conway T, Cohen PS. 2015. Commensal and pathogenic Escherichia colimetabolism in the gut. Microbiol Spectr 3:MBP-0006-2014. https://doi.org/10.1128/microbiolspec.MBP-0006-2014.

78. Kortman GA, Raffatellu M, Swinkels DW, Tjalsma H. 2014. Nutritional ironturned inside out: intestinal stress from a gut microbial perspective.FEMS Microbiol Rev 38:1202–1234. https://doi.org/10.1111/1574-6976.12086.

79. Neidhardt FC, Bloch PL, Smith DF. 1974. Culture medium for enterobac-teria. J Bacteriol 119:736 –747.

80. Wang X, Preston JF, Romeo T. 2004. The pgaABCD locus of Escherichiacoli promotes the synthesis of a polysaccharide adhesin required forbiofilm formation. J Bacteriol 186:2724 –2734. https://doi.org/10.1128/jb.186.9.2724-2734.2004.

81. Baba T, Ara T, Hasegawa M, Takai Y, Okumura Y, Baba M, Datsenko KA,Tomita M, Wanner BL, Mori H. 2006. Construction of Escherichia coli K-12in-frame, single-gene knockout mutants: the Keio collection. Mol SystBiol 2:2006.0008. https://doi.org/10.1038/msb4100050.

82. Bolivar F, Rodriguez RL, Greene PJ, Betlach MC, Heyneker HL, Boyer HW,Crosa JH, Falkow S. 1977. Construction and characterization of newcloning vehicles. II. A multipurpose cloning system. Gene 2:95–113.https://doi.org/10.1016/0378-1119(77)90000-2.

83. Cherepanov PP, Wackernagel W. 1995. Gene disruption in Escherichiacoli: TcR and KmR cassettes with the option of Flp-catalyzed excision ofthe antibiotic-resistance determinant. Gene 158:9 –14. https://doi.org/10.1016/0378-1119(95)00193-a.

84. Timmermans J, Van Melderen L. 2009. Conditional essentiality of the csrAgene in Escherichia coli. J Bacteriol 191:1722–1724. https://doi.org/10.1128/JB.01573-08.

85. Lawhon SD, Frye JG, Suyemoto M, Porwollik S, McClelland M, Altier C.2003. Global regulation by CsrA in Salmonella typhimurium. Mol Micro-biol 48:1633–1645. https://doi.org/10.1046/j.1365-2958.2003.03535.x.

86. Kurabayashi K, Agata T, Asano H, Tomita H, Hirakawa H. 2016. Furrepresses adhesion to, invasion of, and intracellular bacterial communityformation within bladder epithelial cells and motility in UropathogenicEscherichia coli. Infect Immun 84:3220 –3231. https://doi.org/10.1128/IAI.00369-16.

87. Magistro G, Hoffmann C, Schubert S. 2015. The salmochelin receptorIroN itself, but not salmochelin-mediated iron uptake promotes biofilmformation in extraintestinal pathogenic Escherichia coli (ExPEC). Int J MedMicrobiol 305:435– 445. https://doi.org/10.1016/j.ijmm.2015.03.008.

88. Haldimann A, Wanner BL. 2001. Conditional-replication, integration, ex-cision, and retrieval plasmid-host systems for gene structure-functionstudies of bacteria. J Bacteriol 183:6384 – 6393. https://doi.org/10.1128/JB.183.21.6384-6393.2001.

89. Park H, Yakhnin H, Connolly M, Romeo T, Babitzke P. 2015. CsrA partic-ipates in a PNPase autoregulatory mechanism by selectively repressingtranslation of pnp transcripts that have been previously processed byRNase III and PNPase. J Bacteriol 197:3751–3759. https://doi.org/10.1128/JB.00721-15.

90. Vakulskas CA, Pannuri A, Cortés-Selva D, Zere TR, Ahmer BM, Babitzke P,Romeo T. 2014. Global effects of the DEAD-box RNA helicase DeaD(CsdA) on gene expression over a broad range of temperatures. MolMicrobiol 92:945–958. https://doi.org/10.1111/mmi.12606.

Pourciau et al. ®

July/August 2019 Volume 10 Issue 4 e01034-19 mbio.asm.org 18

on Decem

ber 2, 2020 by guesthttp://m

bio.asm.org/

Dow

nloaded from